Aspects of the present invention relate generally to the field of video processing, and more specifically to a predictive video coding system.
In video coding systems, a coder may code a source video sequence into a coded representation that has a smaller bit rate than does the source video and, thereby achieve data compression. A decoder may then invert the coding processes performed by the coder to reconstruct the source video for display or storage.
Frequent exposure adjustments to frames of a video sequence to be coded may lead to inefficient predictive coding since many of the frames may require intra-coding. It can also result in a poor perceptual quality due to the frequent flashing effect resulting from the exposure adjustments. Thus, auto exposure adjustments for every detected luminance change is undesirable. Accordingly, the inventors perceive a need in the art to adapt to changing lighting and image conditions while maintaining proper exposure for frames and minimizing adjustments to the auto exposure features in order to maximize the effectiveness of predictive coding of the captured video.